Overview

This Technical Guide describes the features, installation, configuration and use of the N-Squared Charging Control Services Customer Care Console (N2C5) EDR Analytics Module.

System Introduction

The N2C5 EDR Analytics Module consists of a number of components that ultimately provide business users with access to Intelligent Platform (IN) EDR information.

The data made available through this platform is aggregate data, summarising key numeric data from platform EDRs and CDRs. For example, this system allows quick answers to the following sort of questions:

• How many unique subscribers made calls each day of the past month?

• How do these numbers compare to the previous month, or (if available) previous year?

• What was the $ cost of call charges charged to subscribers for calls made yesterday?

• What was the $ amount of recharges made by subscribers yesterday?

• What is our top called location during the most recent world cup?

• Where in the world is our most called location?

• When last year was our peak call hour?

Data is stored in a PostgreSQL database and is designed to be accessed by business users using business intelligence tools such as Tableau. Data is imported from OCNCC EDRS into this database through an "Extract, Transform and Load" (ETL) process that access EDRs from the OCNCC SMS database.

Data Models

Due to the wide variety of data available through OCNCC EDRs, the analytics module is focused on providing data for two key sets of information:

• Charged call events

• Balance adjustment events

Data Model - Charged Call Events

This data model supports access to the following aggregate information from voice calls made by subscribers:

• Length of calls in seconds.

• Billed Length of calls in seconds. This can differ from actual call length due to billing quanta.

• Total number of calls.

• Total cost of call (in dollars)

• Total number of distinct subscribers who made calls.

Voice call data can be sliced by the following:

• Time of event with hourly granularity. Event time supports week of year as well as the natural Year, Month and Day hierarchy.

• Source location of call. Usually this distinguishes different calling situations within the operator network rather than the actual geographical source.

• Geographical destination called, to the granularity of the clients configuration.

• Type of monetary balance used in call (e.g. promotional cash vs. general cash).

• Type of tariff used in call (if tariffs are used by the platform).

• Subscriber product type at time of call.

Data Model - Balance Adjustment Events

This data model supports access to the following aggregate information from adjustments made to subscriber monetary balance adjustments:

• Number of events.

• Amount (in dollars) of event adjustments.

Data can be sliced by the following:

• Time of event with hourly granularity. Event time supports week of year as well as the natural Year, Month and Day hierarchy.

• Type of monetary balance updated.

• Type of tariff used in call (if tariffs are used by the platform).

• Source of adjustment. This is often the 3rd party system through which the adjustment was made, or the operator who was logged in and performed the adjustment.

• Subscriber product type at time of adjustment.

Non-Functional Considerations

The N-Squared EDR Analytics is built with the following non-functional considerations:

1. The ETL processes are designed to perform quickly enough to process the maximum daily event load as provided for by the license within 24 hours. Access to the resultant data stored in the data models is dependant on the performance of the server.

2. When configured appropriately, it is expected that the data availability lag is less than 24 hours from time of event to the point where it is available in the data models. Processing can be configured to be more frequent or less frequent depending on business user need.

3. All logging by ETL processes is performed to the system logs and may be monitored by standard operations software. Additional statistics are published through statsd and can be viewed in Graphite